Predicting moose behaviors from tri-axial accelerometer data using a supervised classification algorithm

Table 4 Cross-validation confusion matrix for all individuals

Behavior		Predictions								Recall	Precision	Prevalence (%)
Behavior		Lying_u	Ruminating	Foraging	Standing	Lying_o	Walking	Running	Total	Recall	Precision	Prevalence (%)
Observations	Lying_u	120,623	17,762	713	11,612	11,003	378	117	162,208	0.74	0.84	34
	Ruminating	10,499	89,245	603	8649	3948	104	32	113,080	0.79	0.80	24
	Foraging	275	553	84,485	5709	544	7063	78	98,707	0.86	0.90	21
	Standing	11,232	4430	5305	35,571	6304	1973	162	64,977	0.55	0.56	14
	Walking	54	50	2860	624	167	15,647	305	19,707	0.79	0.62	4
	Lying_o	1576	131	56	1416	11,423	30	3	14,635	0.78	0.34	3
	Running	2	18	3	2	1	68	270	364	0.74	0.28	0
	Total	144,261	112,189	94,025	63,583	33,390	25,263	967	473,678

The confusion matrix combines the cross-validation confusion matrixes of the random forest model classifying accelerometer data across all 14 moose observed in captivity in Alaska and Norway. Values in columns represent the number of 3-s accelerometer data intervals predicted for each of the seven behaviors, split into rows based on the behavioral labels of the intervals recorded during the observations. Recall and precision quantify the model classification performance for the respective behavior across all animals in the study. Prevalence indicates the contribution of each behavior to the total sample size of accelerometer data intervals

ISSN: 2050-3385